Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Ghaderi, Susan, Kabgani, Alireza, Moreau, Yves, Ahookhosh, Masoud
Format: Preprint
Veröffentlicht: 2026
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2605.17306
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917505400832000
author Ghaderi, Susan
Kabgani, Alireza
Moreau, Yves
Ahookhosh, Masoud
author_facet Ghaderi, Susan
Kabgani, Alireza
Moreau, Yves
Ahookhosh, Masoud
contents We study sampling from posterior distributions with nonsmooth composite potentials, a setting in which proximal-based Langevin methods are theoretically appealing but in practice limited to simple functions with closed-form proximal operators. We introduce iPULA for composite potentials, an inexact proximal unadjusted Langevin algorithm that replaces exact proximal steps with controlled approximations. Our approach leverages the Moreau envelope to smooth the potential, while allowing inexact evaluation of its gradient through inexact proximal computations. We establish non-asymptotic convergence guarantees for iPULA, explicitly characterizing the impact of inexactness on the sampling error and showing that the inexactness preserves convergence rates up to a quantifiable bias. We demonstrate the practical relevance of iPULA on a medical image reconstruction task, where proximal operators cannot be computed exactly. Experiments demonstrate the effectiveness of iPULA and support our theoretical results.
format Preprint
id arxiv_https___arxiv_org_abs_2605_17306
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Speeding Up Nonsmooth Bayesian MCMC Sampling via Inexact Proximal Unadjusted Langevin Algorithm
Ghaderi, Susan
Kabgani, Alireza
Moreau, Yves
Ahookhosh, Masoud
Optimization and Control
We study sampling from posterior distributions with nonsmooth composite potentials, a setting in which proximal-based Langevin methods are theoretically appealing but in practice limited to simple functions with closed-form proximal operators. We introduce iPULA for composite potentials, an inexact proximal unadjusted Langevin algorithm that replaces exact proximal steps with controlled approximations. Our approach leverages the Moreau envelope to smooth the potential, while allowing inexact evaluation of its gradient through inexact proximal computations. We establish non-asymptotic convergence guarantees for iPULA, explicitly characterizing the impact of inexactness on the sampling error and showing that the inexactness preserves convergence rates up to a quantifiable bias. We demonstrate the practical relevance of iPULA on a medical image reconstruction task, where proximal operators cannot be computed exactly. Experiments demonstrate the effectiveness of iPULA and support our theoretical results.
title Speeding Up Nonsmooth Bayesian MCMC Sampling via Inexact Proximal Unadjusted Langevin Algorithm
topic Optimization and Control
url https://arxiv.org/abs/2605.17306